Hello,
On 2026/01/13 0:44, luigi scarso wrote:
On Sat, 10 Jan 2026 at 07:34, luigi scarso <[email protected]> wrote:
On Thu, 8 Jan 2026 at 09:22, Hironori KITAGAWA via luatex <[email protected]>
wrote:
Hello,
I'm trying to use RTT direction and "Identity-V" encoding fonts
for better vertical typesetting.
But I found that the glyph position is odd:
\hbox dir RTT{%
\HMV % a 10bp font which uses Identity-V CMap
縦\kern5bp 書\kern-10bp き
\vrule height 5bp depth 0pt width 2bp
}
(input: wmode_test.tex (process by luatex-plain);
luatex binary: svn7707)
* glyphs are too low (about the height of "縦") and left (5bp = 0.5em).
* spaces generated by "\kern5bp" and "\kern-10bp" are
reversed in the output.
Is this an intended behavior?
I will check this weekend
Hm perhaps a bad interaction between
res.identity='vertical'; res.writingmode='vertical'
and the RTT dir of hbox
If you comment
res.identity='vertical'; res.writingmode='vertical'
is the
\hbox dir RTT{%
\vrule height 5bp depth 0pt width 2bp%
\HMV %
縦
書
き
\vrule height 5bp depth 0pt width 2bp
}
ok ?
With
> res.identity='vertical'; res.writingmode='vertical'
commented, position of glyphs seems fine.
But they are written by three TJ operators in a PDF,
which makes text extraction from the PDF "incorrect".
BT
/F65 9.999991 Tf
1 0 0 1 72 704.6474 Tm [<094E>]TJ
1 0 0 1 72 692.9274 Tm [<097B>]TJ
1 0 0 1 72 681.6774 Tm [<0356>]TJ
ET
The reason why I want to use
> res.identity='vertical'; res.writingmode='vertical'
is text extraction.
With that, glyphs are written by one TJ operator:
BT
/F65 9.999991 Tf
1 0 0 1 72 704.6474 Tm [<094E>-172<097B>-125<0356>]TJ
ET
Regards
Hironori KITAGAWA
